Mirk Protein Kinase Is Activated by MKK3 and Functions as a Transcriptional Activator of HNF1 (cid:1) *

Mirk/Dyrk1B is an arginine-directed serine/threonine protein kinase that is expressed at low levels in most normal tissues but at elevated levels in many tumor cell lines and in normal skeletal muscle. Colon carcinoma cell lines stably overexpressing Mirk proliferated in serum-free medium, but the mechanism of Mirk action is un-known. DCoHm (dimerization cofactor of hepatocyte nuclear factor 1 (cid:1) ( HNF1 (cid:1) ) from muscle), a novel gene of the DCoH family with 78% amino acid identity to DCoH, was identified as a Mirk-binding protein by yeast two-hybrid analysis and cloned. Mirk co-immunoprecipitated with DCoHm and bound to DCoHm in glutathione S -transfer-ase pull-down assays. DCoH stabilizes HNF1 (cid:1) as a dimer and enhances its transcriptional activity on the (cid:2) -fibrin-ogen promoter reporter, and DCoHm had similar activity. Mirk enhanced HNF1 (cid:1) transcriptional activity in a dose-dependent manner, whereas two kinase-inactive Mirk mutants and a Mirk N-terminal deletion mutant did not. Mirk, DCoHm, and HNF1 (cid:1) formed a complex. Mirk bound to a specific region within the CREB-binding protein-binding region of HNF1 (cid:1) and phosphorylated HNF1 (cid:1) at a site adjacent to the Mirk-binding region. Conversely, the HNF1 (cid:1) binding domain was located within the first five We suggest that subscribers photocopy these corrections and insert the photocopies at the appropriate places where the article to be corrected originally appeared. Authors are urged to introduce these corrections into any reprints they distribute. Secondary services are urged to carry notice of these corrections prominently

Mirk 1 /Dyrk1B is a serine/threonine protein kinase that is expressed at elevated levels in normal skeletal muscle and certain carcinoma cell lines and at low levels in many normal tissues (1). Colon carcinoma cell lines stably overexpressing Mirk proliferated in serum-free medium (1), but the mechanism of Mirk action that enabled this survival capacity is unknown. Mirk is a member of the Dyrk/minibrain family of dual specificity, tyrosine-regulated, arginine-directed protein kinases (2)(3)(4) and is identical to Dyrk1B (5). Mirk/Dyrk1B and the related kinase Dyrk1A exhibit 54% amino acid identity with 90% identity or homology within the conserved kinase domain. Several lines of evidence indicate that Dyrk1A mediates neuronal differentiation. Dyrk1A has been mapped to the Down's syndrome critical region of chromosome 21, overexpression of Dyrk1A has been found in the Down's syndrome fetal brain (6), and transgenic mice overexpressing Dyrk1A exhibit cognitive deficits and motor abnormalities characteristic of Down's syndrome (7). Dyrk1A has been shown to phosphorylate the cAMP-response element-binding protein (CREB) in vivo, leading to the stimulation of subsequent cAMP response element-mediated transcription during neuronal differentiation in hippocampal progenitor cells (8), an activity consistent with its likely mediation of neuronal maturation in vivo. Because Mirk/Dyrk1B was expressed at elevated levels in normal muscle tissue, we screened a muscle cell cDNA library to detect possible Mirk targets and found that Mirk/Dyrk1B activated the transcription factor HNF1␣ (hepatocyte nuclear factor 1␣) and bound to its cofactor DCoH.

EXPERIMENTAL PROCEDURES
Materials-Antibodies to MKK3 and HNF1 were from Santa Cruz Biotechnology, and antibodies to the flag-epitope were from Sigma. Rabbit polyclonal antibody to a unique sequence at the C terminus of Mirk was raised as described (1). Polyvinylidene difluoride transfer paper Immobolin-P was purchased from Millipore. PLUS reagent and LipofectAMINE were from Invitrogen, all radioactive materials were purchased from PerkinElmer Life Sciences, and ECL reagents were from Amersham Biosciences. All other reagents were from Sigma. We thank Stephen E. Mercer of this institution for GST-Mirk preparations.
Cell Culture-NIH3T3 cells and 293T cells were maintained in Dulbecco's modified Eagle's medium containing 7% fetal bovine serum and modified and supplemented as described (9).
Plasmids-pHX9F (Mirk) and pHX9F (kinase-inactive YF Mirk) had been previously generated (1), and all other Mirk expression plasmids, including the pGBKT7-Mirk construct used for the yeast two-hybrid screening, were prepared by Dr. Xiaobing Deng. 2 The (␤-28) 3 -luciferase plasmid encoding three tandem repeats of the ␤-fibrinogen HNF1␣ binding domain in front of a TATA box promoter and a luciferase reporter gene, the expression plasmid pBJ5-DCoH, and the expression plasmid pBJ5-HNF1␣ were the kind gifts of Dr. G. Crabtree, Stanford University. FLAG-MKK3b and MKK3b(E), each in pcDNA3, were the kind gifts of Dr. J. Han, Scripps Institute. A human muscle cDNA library subcloned into the pACT2 vector was purchased form CLONTECH.
Yeast Two-Hybrid Screening-We are grateful to Dr. David Amberg of this institution for help with the two-hybrid screening. For the bait construct, wild-type full-length Mirk cDNA was subcloned into the pGBKT7 vector, because Mirk had no transcriptional activity by itself. The yeast host strain, AH109, contains three different reporter genes that are tightly controlled by the upstream activator sequence, which required GAL4 binding for expression of the reporter gene. AH109 cells were simultaneously co-transformed with pGBKT7-Mirk and the pACT2 muscle cDNA library according to the manufacturer's protocol. The transformants were plated on -His/-Leu/-Trp medium, including 25 mM 3-amino-1,2,4-triasol to screen for the expression of His-3, and incubated at 30°C for 7-10 days. Subsequently, Hisϩ colonies were replated on -Ade/-His/-Leu/-Trp/X-␣-gal medium, and blue colonies were selected as positive candidates. The plasmids from candidate colonies were harvested as described (10). Briefly, cells were grown in 5 ml of selective media overnight and collected by pelleting. The pellets were resuspended in 200 l of lysis buffer and then incubated at 37°C for 30 -90 min after brief vortexing. Lysis buffer consisted of a mixture of 5 ml of SCE buffer, 60 l of 100T zymolyase (10 mg/ml), and 10 l of ␤-mercaptoethanol. Following the addition of 400 l of 0.2 N NaOH/1% SDS solution the lysates were incubated on ice for 5 min and then pelleted, and the supernatants were decanted into fresh tubes and then reclarified by repeated pelleting. The plasmid DNA of each candidate was precipitated and then transformed Escherichia coli DH5␣ using electroporation. The sequences of candidate DNA were analyzed by using an automatic DNA sequencer (model 3700 BC 1.1.0.0, ABI PRISM).
In Vitro Synthesized Mirk and Kinase Reaction-Hexahistidine fusion Mirk was produced by a coupled in vitro transcription and translation system (Promega) using 1 g of pHX plasmid per reaction. Produced protein was immunoprecipitated overnight by using His 6 monoclonal antibody (CLONTECH) and 10 l of 50% protein A-Sephadex (Sigma). The immunoprecipitate was washed three times with radioimmune precipitation buffer and three times with kinase buffer (10 mM Tris, pH7.4, 150 mM NaCl, 10 mM MgCl 2 , 0.5 mM dithiothreitol). The kinase activities of Mirk were tested with either the myelin basic protein (MBP) from Upstate Biotechnology, immunoprecipitated HNF1 overexpressed in 293T cells, or recombinant GST-HNF1 proteins as a substrate according to the protocol by Upstate Biotechnology. Briefly, 10 l of assay dilution buffer (ADB), 10 l of GST-Mirk (200 ng/1 g) in ADB, 10 l of MBP (2 g/l) in ADB, and 10 l of [␥-32 P]ATP (0.5 Ci/l) in a magnesium/ATP mixture (500 M cold ATP and 75 mM MgCl 2 in ADB) were added to the reaction mixture. Immunoprecipitated MKK3E or recombinant p38 (Upstate Biotech.) was added in the mixture. The reaction was performed at 30°C for 15 min with gentle agitation. The reaction was stopped by adding sample buffer and boiling for 5 min. The phospholabeled proteins were separated by SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and detected by autoradiography.
Transient Transfections-293T cells were transiently transfected by adding a complex of LipofectAMINE (2-4 g/g of DNA) in serum-free media for 24 h. For reporter gene assays, NIH3T3 cells were seeded the day before transfection at 0.9 ϫ 10 5 /well and allowed to grow overnight in 7% serum-containing media. Cells were transfected by incubating with a complex of PLUS reagent (3 l/g DNA) and LipofectAMINE (2 l/g DNA) in serum-free media for 18 -24 h in a CO 2 incubator. The amount of total DNA used was kept constant by the addition of empty vector DNA, and luciferase activities were calibrated by co-transfected ␤-galactosidase activity to normalize the transfection efficiency. These assays were carried out in triplicate, and the data shown are representative of three independent experiments.
Co-immunoprecipitations of Mirk and DCoHm (Dimerization Cofactor of HNF1␣ from Muscle)-[ 35 S]-Met-labeled Mirk and DCoHm produced by using the TNT transcription and translation system (Promega) were co-immunoprecipitated in vitro. TNT products (5 l each) were mixed, and 5 l of rabbit polyclonal antibody to the unique C terminus of Mirk or preimmune serum was added with rocking overnight at 4°C. 20 l of a slurry of protein A-agarose conjugates (Santa Cruz Biotechnology) was then added and incubated for an additional 2 h at 4°C. The agarose beads in each tube were extensively washed five times, followed by SDS-PAGE and autoradiography.
Mirk and MKK3E-293T cells in multiple 60-mm dishes were cotransfected with 1 g of the DNA of either MKK3 or MKK3E together with expression plasmids for either Mirk, kinase-inactive YF-Mirk, or pHisA vector (1 g/well), allowed to express for 24 h, and then each dish was lysed in 0.25 ml of EBC buffer (Roche Molecular Biochemicals). An aliquot of total lysate of 300 g was immunoprecipitated with 5 l of an anti-C2 Mirk anti-peptide rabbit polyclonal antibody overnight at 4°C; the complexes were then collected by the addition of 20 l of protein A-agarose, incubated for 2 h at 4°C, washed three times with EBC buffer, and separated by SDS-PAGE.
Immunodetection-Following treatment as indicated and washing with cold phosphate-buffered saline, cells were lysed in EBC or radioimmune precipitation buffer (1ϫ phosphate-buffered saline, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, and the following protease inhibitors: 20 g/ml leupeptin, 20 g/ml aprotinin, 0.5 mM phenylmethylsulfonyl fluoride, 200 M sodium orthovanadate, and 20 mM sodium fluoride). Lysates were pelleted in a microfuge for 15 min to remove insoluble material. Depending upon the experiment, 10 -50 g of cell lysate were blotted onto polyvinylidene difluoride membranes after separation on SDS-PAGE. The blots were blocked in 5% milk in TBST for 1 h at room temperature, incubated for 1 h at room temperature with primary antibody in TBST buffer-3% milk, and proteins were subsequently detected by enhanced chemiluminescence. All Mirk blots used an affinity-purified polyclonal antibody directed to the Mirkunique C terminus. Band density in autoradiograms was measured using a Lacie Silverscanner and Silverscanner III software and analyzed by the IP LabGel program.
GST Pull-down Experiments-Mirk and DCoHm were produced in vitro by a coupled transcription and translation system (Promega) using 1 g of plasmid per reaction. Translation took place in the presence of 2 l of 10 Ci/ml [ 35 S]methionine. The labeled TNT proteins and the GST-fusion proteins were incubated together in binding buffer (20 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM EDTA, pH 8.0, and 0.1% Nonidet P-40) overnight and washed six times in the same buffer before analysis by SDS-PAGE and autoradiography.

Mirk Binds to the HNF1␣ Dimerization Cofactor DCoHm-
To discover the function of Mirk, we used the yeast two-hybrid assay to find Mirk-binding proteins. We isolated a novel gene of the DCoH (11) family, DCoHm, by yeast two-hybrid analysis using full-length Mirk as bait. Yeast two-hybrid could be used because Mirk alone does not transactivate in the yeast twohybrid assay. 3.5 ϫ 10 6 clones of a human skeletal muscle cDNA library were screened to obtain 17 candidate genes, one of which was DCoHm. HNF1␣ is a transcription factor found in endoderm-derived tissues, including kidney, liver, intestine and pancreas, where it may function in maintaining the differentiated phenotype (12). HNF1␣ is also a tissue-specific transcription factor regulating glucose metabolism-related genes in the liver (11), insulin and insulin-like growth factor 1 (IGF-1) (13) in the pancreatic-beta cells, and the protooncogene c-Src in the intestine (14). DCoH binds as a dimer to the unstable HNF1␣ dimer and enables effective binding of the tetrameric complex to DNA (15). DCoH enhances HNF1␣ transcription activity 2-3-fold by stabilizing this complex (11). The DCoH family is highly conserved. DCoHm has 78 and 88% amino acid identity to human and chicken DCoH, respectively ( Fig. 1). DCoHm was considered likely to have activity similar to other members of the DCoH family, because it retained the amino acids necessary for transcriptional activation of HNF1␣, His-62, His-63, and His-80 as well as amino acid Phe-67, which mediates binding to HNF1␣ (16). The interaction of Mirk and DCoHm was confirmed in yeast AH109 cells by a 1.5-2-fold activation of a ␤-galactosidase reporter gene localized downstream of DCoHm (data not shown). Protein interaction was also demonstrated by co-immunoprecipitation of in vitro 35 Slabeled DCoHm with Mirk using antibody to the unique C terminus of Mirk ( Fig. 2) with preimmune serum (P) used in the control. The last lane shows 30% of input DCoHm. The physical association between DCoHm and Mirk shown by coimmunoprecipitation confirms the binding between DCoHm and Mirk in the yeast two-hybrid analysis.
Mirk Enhances the Transcriptional Activity of HNF1␣-Because DCoH functions to stabilize HNF1␣ and increase its transcriptional capacity, we next tested whether Mirk altered the activity of HNF1␣. HNF1␣ has been shown to increase by 60-fold the activity of the reporter construct (␤-28) 3 -Luc, which consists of three copies of the HNF1␣ binding element, whereas DCoH further enhanced this activation about 2-fold (16). We assayed for the effects of Mirk on HNF1␣ transcriptional activation of (␤-28) 3 -Luc in transient transfection assays in NIH3T3 cells, which exhibit low levels of endogenous DCoH. DCoH alone, Mirk alone, or DCoH plus Mirk were unable to activate (␤-28) 3 -Luc. In contrast, the expression of HNF1␣ alone was enough to activate the reporter construct 20-fold, and co-expression of HNF1␣ and DCoH activated the reporter 30-fold (Fig. 3A, first six lanes). Increasing amounts of Mirk were co-transfected with expression plasmids for HNF1␣ and DCoH, which resulted in a dose-dependent activation of HNF1␣, up to five times the activation induced by HNF1␣ and DCoH and 7.5 times the activation by HNF1␣ alone (Fig. 3A, last four lanes). Therefore, Mirk can substantially increase the transcriptional activity of HNF1␣ and may function as a coactivator of HNF1␣ in vivo. What is its significance? Mirk/ Dyrk1B was cloned from colon carcinoma cells, where it is likely to interact with DCoH in vivo. Immunohistochemistry studies demonstrated that DCoH is expressed in each of 20 colon carcinomas independent of Dukes' stage and in each of two colon carcinoma cell lines but not in normal tissue (17), so the presence of DCoH selectively in tumor tissue may allow Mirk through DCoH to activate HNF1␣ in cancer cells.
We next tested whether Mirk might be able to activate HNF1␣ without DCoH. To determine whether DCoH could potentiate the co-activating effect of Mirk, HNF1␣ was transfected at a suboptimal level, one-half of that used in the previous experiment. A 50-fold range of Mirk concentrations was co-transfected with HNF1␣ with or without DCoH (Fig. 3B). Mirk was more effective as a co-activator of HNF1␣ in the presence of DCoH. The highest Mirk concentration tested induced over twice as much reporter gene activity in the presence of DCoH as in its absence. Mirk, like DCoH and HNF1␣, forms dimers (data not shown). These data together suggest that HNF1␣, DCoH, and Mirk could act as a heterohexamer, with dimers of DCoH stabilizing the HNF1␣ dimer and serving as an attachment factor for dimerized Mirk.
Mirk is a serine/threonine protein kinase activated by autophosphorylation on tyrosine (1). Was the capacity to phosphorylate HNF1␣ necessary for the activation of this transcription factor by Mirk? Wild-type Mirk increased HNF1␣ activation 3-fold when co-transfected with DCoH and HNF1␣, whereas the co-transfection of two Mirk kinase-inactive mutant constructs, Mirk-YF and Mirk-KR (1), had no effect (Fig. 4A), demonstrating that the kinase activity of Mirk is essential for its ability to increase HNF1␣ function. Further mutational analysis of Mirk was performed. Mirk is composed of a central conserved kinase domain flanked by non-conserved N-terminal and C-terminal sequences (1). Mirk deletion constructs were made; ⌬NMirk consists of amino acids 110 -629, and ⌬CMirk consists of amino acids 1-435. The deletion of the N-terminal region of Mirk completely suppressed the activation of HNF1␣ by Mirk, whereas the deletion of the Mirk C terminus had little effect (Fig. 4A). We next tested whether a double mutant, transcriptional activation in a dose-dependent manner in the absence of DCoH, whereas kinase-inactive YF-Mirk and the Mirk deletion mutant ⌬NMirk had no activity (Fig. 4B). These data indicate that Mirk kinase activity was essential for Mirk function as a transcriptional co-activator of HNF1.
Synergism between Mirk and MKK3E-Transcription factors are often activated through a kinase cascade such as the well known MAP kinase superfamily. Mirk was shown to be a substrate of extracellular signal-related kinase 2 (ERK2), c-Jun NH2-terminal kinase (JNK1), and p38 in vitro (1), suggesting that one or more MAP kinases might activate Mirk in vivo, allowing Mirk to then activate HNF1␣. We screened a series of MAP kinase kinases and their activators for their ability to substitute for Mirk in HNF1␣ activation using the ␤-fibrinogen promoter reporter construct (␤-28) 3 -Luc in transient transfection experiments. The MAP kinase kinase MKK3, an activator of p38 (18 -20), was capable of activating HNF1␣ (Fig. 5). The constitutively activated form of MKK3, doubly mutated at its activation domain MKK3E (S189E/T193E) (21), was used in place of MKK3 because it activates its downstream kinase targets without itself requiring activation by upstream signals. Both Mirk and MKK3E activated HNF1␣ 3-5-fold. When Mirk and MKK3E were added together, they induced a synergistic increase of 15-fold HNF1␣ activation. Activation of HNF1␣ as a transcription factor, potentiated by MKK3E, also required the same Mirk functions as described earlier (Fig. 4), i.e. Mirk kinase activity, the nonconserved N terminus of Mirk, but not the nonconserved C terminus of Mirk. Wild-type Mirk could not be replaced by either kinase-inactive Mirk or ⌬NMirk but could be replaced to some extent by ⌬CMirk (Fig. 5). MKK3E is a potent activator of p38 (21). These data suggested three models. 1) Mirk and MKK3E activated HNF1␣ independently. 2) MKK3E directly activated Mirk by phosphorylation. 3) MKK3E activated p38, which in turn phosphorylated and activated Mirk.
MKK3E Binds to Mirk and Activates It-To test this hypothesis, we initially determined whether Mirk and MKK3 interacted directly by co-immunoprecipitation experiments. 293T cells were co-transfected with either MKK3 or MKK3E together with either wild-type Mirk, kinase-inactive YF-Mirk, or epitope-tagged vector, and Mirk was immunoprecipitated after expression (Fig. 6A). A distinct band of MKK3E was observed in both the immunoprecipitates of wild-type Mirk and kinaseinactive Mirk. Much less (6 -7%) wild-type MKK3 associated with Mirk, although similar amounts of MKK3E, MKK3, and Mirk were synthesized in the lysates (Western blot of total lysates shown in lower two panels, Fig. 6A). Therefore, MKK3E associated directly with Mirk, and any effect MKK3E exerted on HNF1␣ could be mediated through exogenous or endogenous Mirk (Fig. 5, lanes 2 and 3, respectively). Because MKK3 is a potent activator of p38 (21), it was also possible that MKK3 activates Mirk. MKK3E was expressed in 293T cells, immunoprecipitated, and added to in vitro kinase reaction mixtures containing recombinant purified Mirk or recombinant purified p38, as indicated, with [␥-32 P]-ATP and MBP as substrates. The reaction mixtures containing [␥-32 P]-ATP were analyzed by SDS-PAGE and autoradiography (Fig. 6B). Mirk phosphorylation of MBP was increased ϳ50% in the presence of MKK3E, although MKK3E itself did not phosphorylate MBP. As a control, MKK3E was shown to increase the activity of a small concentration of p38 as an MBP kinase ϳ50% (Fig. 6B,  last two lanes). Therefore, MKK3 binds to Mirk and increases Mirk kinase activity. Because MKK3 mediates various environmental stress signals, it is possible that MKK3 increases the function of Mirk as a transcriptional activator of HNF1␣ in response to certain stresses. Stably overexpressed Mirk enables cells to proliferate and remain viable in serum-free conditions (1), possibly because Mirk activates the transcription of survival factors such as insulin-like growth factor 1 through the activation of HNF1␣.
Mirk Binds to a Specific Region within the CBP-binding Region of HNF1␣-We next determined the regions of Mirk and HNF1␣ necessary for their interaction. Glutathione Stransferase pull-down assays confirmed a three-part association between DCoHm, Mirk, and HNF1 (Fig. 7A, lanes 4 -6). Both DCoHm and Mirk were 35 S-labeled by coupled transcription and translation reactions, incubated with either GST-HNF1␣ or GST coupled to beads, and the bead eluates were examined by PAGE followed by autoradiography. DCoHm and Mirk, when added individually or added together, bound to HNF1␣ (Fig. 7A). DCoHm bound more strongly to HNF1␣ than did Mirk when the bound fraction to total input was compared. There was no additive effect of DCoHm and Mirk in multiple experiments. Neither DCoHm or HNF1␣ bound to GST.
Next, a series of deletion mutants of HNF1␣ were generated, and GST-pull down assays were repeated with 35 S-labeled Mirk (Fig. 7B). Mirk bound to HNF1␣ fragments encompassing amino acids 1-203 and 1-283 but not to fragments of the N terminus (amino acids 1-111) or C-terminal regions (amino acids 283-481 and 481 to 628). Therefore, the minimal region of Mirk/HNF1␣ binding encompassed HNF1␣ amino acids 112-203. This region was similar to the minimal region for HNF1␣ interaction with the co-activator CBP (22), amino acids 95-295. We next determined which region of Mirk was necessary for binding to HNF1␣ by performing GST pull-down assays with deletion constructs of Mirk (Fig. 8). The constructs consisted of ⌬NMirk (amino acids 110 -629) cleaved of the nonconserved N terminus; ⌬N2Mirk (amino acids 233-629) cleaved of the nonconserved N terminus and the first five subdomains of the conserved kinase domain and the ATP binding site K140; ⌬233-435, a partial kinase-domain deletion of amino acids 233-435 that retained the N and C termini; ⌬CMirk (amino acids 1-435 with the nonconserved C terminus deleted); and C, an unrelated control protein p53. There were equivalent amounts of Mirk deletion products added to the GST pull-downs (Fig. 8, lower panel). However, ⌬N2 Mirk bound very poorly to GST-HNF1␣-(1-203) compared with the other deletion constructs and to full-length Mirk (Fig. 8, lane 1). Therefore, the Mirk region of amino acids 111-233 encompassing the first five conserved kinase subdomains was essential for Mirk binding to HNF1␣.
Mirk Phosphorylates HNF1␣-A direct interaction between Mirk and HNF1␣ occurred in the absence of DCoHm, shown in the GST pull-down assays (Fig. 7A, lane 5). These results suggested that DCoHm simply facilitated Mirk association with HNF1␣ and that HNF1␣ might be the immediate target of Mirk. This was shown to be the case, because Mirk phosphorylated GST-HNF1␣ in vitro in the absence of DCoHm while exhibiting no kinase activity on GST itself (Fig. 9A) or on DCoHm (data not shown). Therefore, Mirk binds to DCoHm in the DCoHm/HNF1␣ tetramer and is thus able to directly bind and then phosphorylate HNF1␣. There was no canonical arginine-directed Dyrk1A phosphorylation site (4) within the region of HNF1␣ that binds to Mirk (HNF1␣, amino acids 112-203). However, a perfect Dyrk1A phosphorylation site was seen at Ser-249, which is still within the CBP binding domain (22). Mutation of this site decreased phosphorylation by Mirk 2.5fold (Fig. 9B). Mirk binds to a region within the N terminus of HNF1␣ and then phosphorylates an amino acid just distal to the binding region but still within the CBP binding domain. Additional studies will determine whether phosphorylation at this site increases CBP binding and HNF1␣ activation in vivo. DISCUSSION Mirk/Dyrk1B, like its family member Dyrk1A, is now shown to function as a transcription factor activator. There is strong evidence that Dyrk1A functions in neurogenesis and brain development (6,7,23,24). Dyrk1A activity was induced during the differentiation of immortalized hippocampal progenitor cells, and the addition of the neurogenic factor basic fibroblast growth factor to these cells resulted in the specific binding of Dyrk1A to CREB, the phosphorylation of CREB by Dyrk1A, and the stimulation of CRE-mediated gene transcription (8). Mirk/Dyrk1B is also expressed in the brain, but its highest expression is in skeletal muscle (1). Screening a normal skeletal muscle cDNA library led us to discover that Mirk binds to a novel member of the highly conserved DCoH family, DCoHm. DCoH binds as a dimer to the unstable HNF1␣ dimer and enables effective binding of the tetrameric complex to DNA (15). DCoH enhances HNF1␣ transcription activity 2-3-fold by stabilizing this complex (11), and Mirk increased this activity a further 5-fold. The interaction between DCoHm and Mirk was confirmed by co-immunoprecipitation studies and GST-pull down assays. Mirk was shown to substantially increase the transcription activity of HNF1␣ in transient transfection assays and may function as a co-activator of HNF1␣ in vivo. Mirk did not require DCoH to active HNF1␣. Mirk, through the N terminus of its conserved kinase domain, bound to HNF1␣ at a site within its CBP binding domain and then directly phosphorylated HNF1␣ adjacent to the binding site but still within the CBP binding domain. Mirk kinase activity was required for transcriptional activation of HNF1␣, because kinase-inactive mutants were also unable to activate HNF1␣.
HNF1␣ mediates the transcription of several genes that potentially could contribute to the differentiation and growth of normal muscle cells through Mirk interaction with DCoHm. However, Mirk is expressed in several cell lines established from solid tumors, including colon, lung, and ovarian, with much less expression in leukemia and lymphoma-derived cell lines (1). DCoH and Mirk are both expressed in colon carcinomas. DCoH was not found in normal colon tissue but was found in every colon carcinoma examined by immunohistochemistry (17), and Mirk was found in each of seven colon carcinoma cell lines (1). Mirk, DCoH, and HNF1␣ were found to form a complex in vitro, thus Mirk may complex with the DCoH/HNF1␣ tetramer in vivo. Therefore, in many colon carcinomas Mirk and DCoH are co-expressed and may function as an activating complex for HNF1␣ to induce ectopic gene expression. The expression of some of these genes may contribute to the ability of cell lines with stably overexpressed Mirk to maintain serumfree proliferation (1).
Recombinant Mirk is a constitutively active kinase. However, its kinase activity on MBP and its transcriptional activation of HNF1␣ was enhanced by MKK3, an upstream activator of the stress-activated MAP kinase p38 (21). MKK3 also directly bound to Mirk in vivo as demonstrated in co-immunoprecipitation experiments. MKK3 is activated by phosphorylation at Ser-189 and Thr-193 within a P-activation loop of its conserved kinase subdomain VIII (21) by upstream kinases in response to stress agents. p38, in turn, is activated by dual phosphorylation at threonine and tyrosine within the TPY motif in conserved kinase subdomain VIII (25)(26)(27). However, Mirk/Dyrk1B's activation domain is YQY, which is at a position located within conserved subdomains VII and VIII of the cat-alytic domain homologous to the p38 activation domain (1). Recent studies of the related kinase Dyrk1A have shown that only the second tyrosine in the activation domain, Tyr-321 and not Tyr-319, was autophosphorylated, and the mutation of Tyr-319 to Phe failed to reduce kinase activity (28). MKK3 possibly activates Mirk by phosphorylating it at the second tyrosine in its activation domain. Autophosphorylated tyrosine residues were also found in the N terminus of Dyrk1A, and deletion of the N terminus suppressed kinase activity (28) as deletion of the N terminus of Mirk suppressed its transactivator activity (Fig. 4). Dyrk1A is an arginine-directed kinase (4), and comparison of the Dyrk1A and ERK2 catalytic cores revealed that Tyr-321 of Dyrk1A can interact with Arg-325 and Arg-328 (28). There are also two comparable arginine residues upstream of the YQY activation motif of Mirk. Although Dyrk1A and Mirk/Dyrk1B can autophosphorylate in bacteria, greater phosphorylation was seen when both proteins were expressed in mammalian cells. MKK3 binding sites and phosphorylation sites on Mirk must be mapped to uncover the precise relationship between MKK3 and Mirk.